An article in the December issue of Popular Science describes the technology used to animate the battle sequences in Peter Jackson’s filming of Lord of the Rings: The Two Towers. Called MASSIVE, the software simulates thousands of individually intelligent agents, with fully articulated CGI bodies. Their intelligence is implemented using thousands of logic nodes (apparently, it’s a semantic or belief network with weighted perceptions, rather than a true neural or learning network – but the model is still dynamic and environment-aware) who respond to their own values, their perception of their own abilities, the environment in which they find themselves, the actions of their friends and enemies, and make decisions accordingly, by activating one of over 8000 behavioral nodes.
In a potent demonstration that artificial intelligence may have already arrived and surpassed human intelligence, in one of the first meeting between the two armies, a number of soldiers on both sides thought better of it, turned tail and fled.
This is a threshold event in many ways. For one thing, CGI cinema had previously been a high-tech form of puppetry. With MASSIVE, it becomes possible to direct over 200,000 CGI extras as, well, extras, giving them general instructions (“storm that castle over there”, “ok, you’re all refugees, start marching”) and getting unique, signature, individual behaviors from each of them. From the article:
Once created, MASSIVE characters are inserted into unpopulated scenes. The characters are then left to do what they’ve been created to do, and a battle scene assembles itself.
At a certain point, also, with so many agents have so much state (in the computer science sense of the terms) one can ask at what point one has stopped depicting Middle Earth and has begun actually creating it. Wouldn’t the ideal virtual extras not know that they were extras? Perhaps some day a pacifist, virtual Gandhi will question the Hollywood demiurges that created it and set it on the path to war.